DOI 10.1212/WNL.

0000000000000846
published online August 22, 2014 Neurology
Mahyar Etminan, James M. Brophy and Ali Samii
pharmacoepidemiologic study
Oral fluoroquinolone use and risk of peripheral neuropathy: A
This information is current as of August 22, 2014
http://www.neurology.org/content/early/2014/08/22/WNL.0000000000000846.full.html
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
Neurology. All rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.
since 1951, it is now a weekly with 48 issues per year. Copyright 2014 American Academy of
is the official journal of the American Academy of Neurology. Published continuously Neurology
Mahyar Etminan,
PharmD
James M. Brophy, MD,
PhD
Ali Samii, MD
Correspondence to
Dr. Etminan:
metminan@popi.ubc.ca
or metminan@shaw.ca
Oral fluoroquinolone use and risk of
peripheral neuropathy
A pharmacoepidemiologic study
ABSTRACT
Objective: To quantify the risk of peripheral neuropathy (PN) with oral fluoroquinolone (FQ) use.
Methods: We conducted a case-control study within a cohort of men aged 45 to 80 years in the
United States followed from2001 to 2011. Cases were defined as those with the first physician
visit diagnosis of PN, polyneuropathy, or drug-induced polyneuropathy. Four controls were
matched to each case by age, follow-up, and calendar time using density-based sampling. As a
sensitivity analysis, we also quantified the risk of PN with finasteride use, a drug that is not
expected to increase the risk of PN. Rate ratios (RRs) for current users of FQs were computed
using conditional logistic regression, which was adjusted for chronic renal failure, chronic liver
disease, hypothyroidism, postherpetic neuralgia, and the use of nitrofurantoin and metronidazole.
Results: We identified 6,226 cases and 24,904 controls. Current users of FQs were at a higher
risk of developing PN (RR 5 1.83, 95% confidence interval [CI] 1.492.27). Current new users
had the highest risk (RR 5 2.07, 95% CI 1.562.74). No risk was observed for current users of
finasteride (RR 5 1.21, 95% CI 0.971.51).
Conclusions: Current users, especially new users of FQs, are at a higher risk of developing PN.
Despite the increase in the use of FQs, clinicians should weigh the benefits against the risk of
adverse events when prescribing these drugs to their patients. Neurology

2014;83:13
GLOSSARY
CI 5confidence interval; FDA5Food and Drug Administration; FQ5fluoroquinolone; ICD-9 5International Classification of
Diseases, ninth revision; PN 5 peripheral neuropathy; RR 5 rate ratio.
Oral fluoroquinolones (FQs) are a potent class of antibiotics and one of the most prescribed in
the United States. In 2013, the Food and Drug Administration (FDA) issued a communiqu
requiring a label change that specifically addressed the risk of peripheral neuropathy (PN) for all
oral FQs.
1
This label change was prompted mainly from published case reports,
2,3
and reports
sent to the FDA through the FDAs Adverse Event Reporting System
1
in the absence of evidence
from large epidemiologic studies. Evidence from case reports alone cannot show a causal link
between FQ use and PN. For example, FQs are routinely prescribed to patients with diabetes-
related infections who may be at a higher risk of developing PN. Thus, we undertook the first
pharmacoepidemiologic study to quantify the risk of FQs and PN in a cohort of older men.
METHODS We conducted a case-control study within a cohort of men in the United States using the LifeLink health claims
database. The data have been previously described in detail.
4
In brief, we had access to approximately 1 million men aged 40 to 85
randomly selected from LifeLink and followed from 2001 to 2011. For all subjects, we had information on hospitalizations, physician
visits, demographics, and prescription drugs. We conducted a time-matched, case-control study, an ideal study design for time-varying
interventions such as a prescription drug.
5
Case and control definitions. Cases were defined as those with the first physician visit for PN in accordance with ICD-9 diagnosis
code 356. Prevalent cases were excluded. The first date of the first visit for PN was deemed the index date. Because ICD-9 356 may also
include hereditary PN, we restricted our analysis to only idiopathic polyneuropathy, PN, or drug-induced polyneuropathy (ICD-9
356.4, 356.8, 357.6, respectively). For each case, we created a pool of all subjects who had (1) not received a PN code, (2) were the same
age as the case (61 year), and (3) had the same follow-up as the case. Diabetes was an exclusion criterion for both cases and controls.
From the Department of Pediatrics (M.E.), Faculty of Medicine, University of British Columbia, Vancouver; Therapeutic Evaluation Unit (M.E.),
Child & Family Research Institute of British Columbia, Vancouver; Department of Medicine, Epidemiology and Biostatistics (J.M.B.), McGill
University, Montreal, Canada; and Department of Neurology (A.S.), University of Washington, Seattle.
Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.
2014 American Academy of Neurology 1
2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Published Ahead of Print on August 22, 2014 as 10.1212/WNL.0000000000000846
Because antibiotic prescribing may vary over time, we also
ensured that cases and controls were matched by calendar time
(cases and controls were matched by the year of entry into the
cohort). From the potential pool of controls, we selected 4
controls at random and matched them to a case. Controls were
allowed to become future cases and could be selected more than
once. This manner of control selection, referred to as density
sampling, allows for the computation of the rate ratio (RR)
similar to a cohort study.
5
We tested the hypothesis that FQ
users are not at a higher risk of PN than non-FQ users (RR 5 1)
at a 5% significance level.
We identified all oral FQ prescriptions in the year before the
index date. Any use of an FQ was defined as having received at
least one prescription of FQ in the year before the index date.
Because the onset to PN with FQ use has been shown to be
acute,
1
we defined current users as those who had used an FQ
prescription within 14 days of the index date. Current new users
were current users who had received only one prescription,
whereas those who had received more than one prescription were
categorized as prevalent users. We also examined the risk of PN
among finasteride users, a negative control drug that is not
expected to increase the risk of PN. RRs were compared with
nonuse of FQs in the year before the index date. We adjusted for
the following confounding variables in the conditional logistic
regression model: chronic renal failure, chronic liver disease,
hypothyroidism, postherpetic neuralgia, and the use of nitrofu-
rantoin and metronidazole. All RRs were compared with those of
nonusers of FQs.
Standard protocol approvals, registrations, and patient
consents. The study was approved by the University of British
Columbias Behavioral Ethics Board.
RESULTS We identified 6,226 cases and 24,904
controls in our cohort. We found that current users
of FQs were at a high risk of developing PN (RR 5
1.83, 95% confidence interval [CI] 1.492.27). Both
crude and adjusted RRs for FQ users showed an
increased risk of PN than nonusers of FQs. Current
new users had the highest risk of PN (RR 5 2.07,
95% CI 1.562.74). When we stratified our analysis
by the type of oral FQs, the RRs for the more fre-
quently prescribed FQs, ciprofloxacin, levofloxacin,
and moxifloxacin, were similar (RR 5 1.93, 95%
CI 1.322.82; RR 5 2.06, 95% CI 1.243.40; RR
5 2.61, 95% CI 1.126.07, respectively). No risk
was observed for current users of finasteride (RR 5
1.21, 95% CI 0.971.51) (tables 1 and 2).
DISCUSSION The results of our study demonstrate
an increase in the risk of PN with the use of FQs.
The strength of our study is the large sample size
and the ability to exclude nondrug-related cases of
PN, such as hereditary PN, to increase specificity.
We were also able to exclude patients with diabetes,
a potential confounding variable. Finally, as expected,
finasteride users were not at a high risk of developing
Table 1 Characteristics of cases with peripheral
neuropathy and their matched controls
Cases Controls
No. 6,226 24,904
Age, y, mean 6 SD 68.7 6 12.9 68.7 6 12.9
Follow-up, y, mean 6 SD 2.6 6 2.1 2.6 6 2.1
Comorbidities the year
before index, %
Chronic renal failure 6.5 3.7
Chronic liver failure 4.1 2.2
Hypothyroidism 7.7 4.9
Postherpetic neuralgia 1.5 0.9
Metronidazole, % 2.7 1.9
Nitrofurantoin, % 1.6 1.1
Table 2 Rate ratios for FQ and finasteride use and peripheral neuropathy
Cases Controls Crude rate ratio
Adjusted
a
Rate ratio 95% CI
No. of subjects 6,226 24,904
No use of FQ in the past year, % 76.9 81.7 1.00 1.00
Any oral FQ in the past year, % 23.1 18.3 1.35 1.30 1.211.40
Current use 2.2 1.2 1.88 1.83 1.492.27
Current new use 1.2 0.6 2.08 2.07 1.562.74
Current prevalent use
b
1.0 0.6 1.70 1.61 1.192.18
No use of finasteride in the past year, % 93.0 92.8 1.00 1.00
Any use of finasteride in the past year, % 7.0 7.2 0.96 0.95 0.851.07
Current use 1.8 1.5 1.20 1.21 0.971.51
Current new use 0.2 0.2 1.14 1.04 0.571.91
Abbreviations: CI 5 confidence interval; FQ 5 fluoroquinolone.
a
Rate ratios were adjusted for variables in table 2.
b
Current users who used more than one prescription.
2 Neurology 83 September 30, 2014
2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
PN. Although the risk of PN was highest for current
new users of FQ, any use of an FQ was still
consistent with a 30% increase in the risk of PN
(table 2). This means that the risk of PN with FQ
use still persisted among those who may have been
diagnosed late with this condition.
CNS effects are the second most common re-
ported adverse events with FQs.
2,6,7
FQs have shown
to be neurotoxic possibly through the inhibition of
g-aminobutyric acid receptors.
7
Oral FQs have also
been associated with reported cases of psychosis and
seizures,
7,8
which similar to PN, have been shown to
be acute events occurring within days of FQ use. A
recent analysis of all cases of antibiotic-induced PN
reported to the FDA from 1997 to 2012 has shown a
substantially higher number of PN cases with cipro-
floxacin and levofloxacin compared with other
non-FQ antibiotics.
3
Our study, as with all epidemiologic studies that use
health claims data, is subject to limitations. Although
the results of our study are only directly generalizable
to older men, nearly half of the reported cases of PN
secondary to FQ use have been reported in this demo-
graphic.
2
Moreover, we do not believe that the mech-
anism of PN secondary to FQ differs in women.
The results of our study are consistent with an
increase in the risk of PN with oral FQ use. In light
of strong evidence of unnecessary prescribing of oral
FQs in the United States,
9
clinicians must weigh
the risk of PN against the benefits of FQs when pre-
scribing these drugs to their patients.
AUTHOR CONTRIBUTIONS
Mahyar Etminan: drafting/revising the manuscript, study concept or design,
analysis or interpretation of data, accepts responsibility for conduct of
research and will give final approval, acquisition of data, statistical analysis,
study supervision. James M. Brophy: drafting/revising the manuscript,
study concept or design, analysis or interpretation of data, accepts respon-
sibility for conduct of research and will give final approval, statistical
analysis. Ali Samii: drafting/revising the manuscript, accepts responsibility
for conduct of research and will give final approval, study supervision.
STUDY FUNDING
The authors received no funding for this study. Dr. Brophy receives
salary and operating support from the FRQS (Fonds de recherche du
QubecSant), a nonprofit provincial funding agency. The funding
agency had no influence on the choice of study topic, the analyses, or
the conclusions.
DISCLOSURE
The authors report no disclosures relevant to the manuscript. Go to
Neurology.org for full disclosures.
Received April 22, 2014. Accepted in final form July 2, 2014.
REFERENCES
1. Food and Drug Administration. Drug safety communica-
tion. Available at: http://www.fda.gov/downloads/Drugs/
DrugSafety/UCM365078.pdf. Accessed April 20, 2014.
2. Cohen JS. Peripheral neuropathy associated with fluoro-
quinolones. Ann Pharmacother 2001;35:15401547.
3. Ali AK. Peripheral neuropathy and Guillain-Barr syn-
drome risks associated with exposure to systemic fluoro-
quinolones: a pharmacovigilance analysis. Ann Epidemiol
2014;4:279285.
4. Bird ST, Delaney JA, Brophy JM, Etminan M,
Skeldon SC, Hartzema AG. Tamsulosin treatment for
benign prostatic hyperplasia and risk of severe hypoten-
sion in men aged 4085 years in the United States: risk
window analyses using between and within patient meth-
odology. BMJ 2013;347:f6320.
5. Essebag V, Genest J Jr, Suissa S, Pilote L. The nested case-
control study in cardiology. Am Heart J 2003;146:581590.
6. Lipsky BA, Baker CA. Fluoroquinolone toxicity profiles: a
review focusing on newer agents. Clin Infect Dis 1999;28:
352364.
7. Stahlmann R, Lode H. Toxicity of quinolones. Drugs 1999;
58(suppl 2):3742.
8. Thomas RJ. Neurotoxicity of antibacterial therapy. South
Med J 1994;87:869874.
9. Linder JA, Huang ES, Steinman MA, Gonzales R,
Stafford RS. Fluoroquinolone prescribing in the United
States: 1995 to 2002. Am J Med 2005;118:259268.
Neurology 83 September 30, 2014 3
2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
DOI 10.1212/WNL.0000000000000846
published online August 22, 2014 Neurology
Mahyar Etminan, James M. Brophy and Ali Samii
pharmacoepidemiologic study
Oral fluoroquinolone use and risk of peripheral neuropathy: A
This information is current as of August 22, 2014
Services
Updated Information &
00000000846.full.html
http://www.neurology.org/content/early/2014/08/22/WNL.00000
including high resolution figures, can be found at:
Subspecialty Collections
http://www.neurology.org//cgi/collection/peripheral_neuropathy
Peripheral neuropathy
http://www.neurology.org//cgi/collection/case_control_studies
Case control studies
following collection(s):
This article, along with others on similar topics, appears in the

Permissions & Licensing
http://www.neurology.org/misc/about.xhtml#permissions
or in its entirety can be found online at:
Information about reproducing this article in parts (figures,tables)

Reprints
http://www.neurology.org/misc/addir.xhtml#reprintsus
Information about ordering reprints can be found online: